Most DSM-III-R personality disorders do not have strong empiric support. In addition, objections have been raised concerning the categorical system of diagnosis employed in the DSM-III-R. While self rating scales have traditionally been used in personality assessment, structured interviews have been suggested as a means to enhance the reliability of DSM-III-R personality disorder diagnoses. While establishing the validity of such diagnoses is a formidable task, the prediction of treatment response has been suggested as validator of personality disorder diagnosis . The goals of the present study are: 1) to examine the correspondence between personality assessments based on DSM-III-R criteria with those obtained using a well established instrument based on a dimensional model of personality and: 2) to determine the reliability and relative prognostic utility of these personality assessments. Patients in a substance abuse program will be assessed using a structured interview and a self report questionnaire, both designed to yield DSM- III-R personality disorder diagnoses. They will also complete the Eysenck Personality Questionnaire (EPQ). Addiction severity will be assessed using the Addiction Severity Index (ASI, McClellen et al, 1980). A six- month follow-up assessment will be obtained using the same instruments. The long-term objectives are to assess the reliability and validity of DSM-III-R personality disorder diagnoses, and to examine the postulated correspondence between the DSM-III-R personality disorder system and a well established dimensional model.